Chapter 72. Neuromodulation for Pain

In the past 30 years, the field of pain management has increasingly incorporated technologies of neurostimulation as part of the treatment algorithm for patients with intractable pain. These technologies include peripheral nerve stimulation, spinal cord stimulation, deep brain stimulation, sacral nerve stimulation, and trigeminal nerve stimulation. More and more patients with complex pain conditions who have failed more conservative management are finding some degree of relief through the use of these devices.

The inspiration for development of these technologies came from the landmark “gate control theory” introduced by Melzack and Wall in 1965.1 Although this model fails to explain certain phenomena seen in painful conditions and cannot account for all of the observed effects of neurostimulation, the gate control theory remains the primary paradigm used to describe how neurostimulation acts to modify pain transmission. The gate control theory is based on the presence of cells in the dorsal horn of the spinal cord that receive afferent signals from peripheral C fibers conveying painful stimuli, as well as non-nociceptive sensory fibers. Once pain signals reach these dorsal horn interneurons, a “gate” is activated and painful impulses propagate along ascending fibers to the brain, resulting in conscious awareness of pain. Wall and Melzack proposed that the gate could be closed to the transmission of painful impulses through the selective activation of non-nociceptive fibers. Thus, the notion of using neurostimulatory devices to preferentially activate non-nociceptive fibers as a means of diminishing pain was born.

The application of neurostimulation has been increasing since C. Norman Shealy implanted the first spinal cord stimulator device in 1967. The scope of conditions amenable to these technologies has likewise expanded as they have become more widely used in varying populations of patients around the world. Current indications for the use of these devices include isolated peripheral nerve injuries, failed back surgery syndrome, peripheral vascular disease with critical limb ischemia, refractory angina pectoris, deafferentation syndromes, spinal-cord-injury–related pain, interstitial cystitis, and trigeminal neuralgia. Other applications are continuously being evaluated, including the use of spinal cord stimulation as a means to monitor evoked potentials during thoracoabdominal aneurysm repair.2

Although the gate control theory seems to explain the primary mechanism by which neurostimulation works, it cannot account for some of the clinically observed effects of spinal cord stimulation. Some researchers have proposed that spinal cord stimulation inhibits transmission of painful impulses partly by inducing a differential conduction block of afferent nociceptive fibers via antidromic stimulation. That the effects of stimulation can outlast the duration of the stimulation would seem, however, to indicate this is not the only mechanism by which neurostimulation influences pain transmission. A number of investigators have speculated that neurohumoral mechanisms are also involved. Studies have been conducted to further elucidate which mediators may contribute. Substances that have been purported to be involved in the neuromodulatory effects of spinal cord stimulation include endogenous opioids, gamma (γ)-aminobutyric acid (GABA), adenosine, substance P, serotonin, calcitonin gene-related peptide (CGRP), and nitric oxide. Some of these mediators may also play a role in the sympathetic inhibitory effects of spinal cord stimulation.

Endogenous Opioids

It has been postulated by some researchers that endogenous opioid release may occur as a result of neurostimulation. There does exist evidence that transcutaneous electrical nerve stimulation (TENS) results in increased cerebrospinal fluid (CSF) levels of several types of opiate peptides. Some of these peptides appear to be μ-receptor specific, whereas others are more κ specific. The type of peptide released in response to TENS seems to depend on the frequency of the stimulation. The pain-relieving effects of low-frequency TENS appear to be reversible with naloxone, whereas those of high-frequency TENS are not.3 There is no consistent evidence that endogenous opioid release occurs as a result of spinal cord stimulation, however; and naloxone does not reverse the effects of spinal cord stimulation. Thus, while endogenous opioid release may result from some forms of neurostimulation, it does not appear to play a significant role in spinal cord stimulation.

GABA

GABA is a major inhibitory neurotransmitter that has been shown to mediate some of the pain-relieving effects of spinal cord stimulation. GABA is able to modulate pain by controlling the release of excitatory amino acids such as aspartate and glutamate in the dorsal horn. It is believed that certain anomalies seen in chronic pain states, such as allodynia and hyperalgesia, may be caused by disrupted regulation of excitatory amino acid release in the dorsal horn.4 This is normally kept in check by local GABA activity. Deficient GABA function as a consequence of nerve injury may result in abnormal responsiveness of wide-dynamic range neurons to tactile stimulation. The sequela of this may be the development of allodynia and hyperalgesia.

It was speculated that some of the antinociceptive actions of spinal cord stimulation are due to the ability of spinal cord stimulation to modulate GABA release in the dorsal horn. Many studies have been done to evaluate this hypothesis. In one of these studies, investigators measured GABA levels in the dorsal horn before and after initiation of spinal cord stimulation in laboratory animals exhibiting the behavioral equivalent of allodynia in humans. It was revealed that the animals that responded to the stimulation with diminished withdrawal responses to tactile stimuli indeed had enhancement of dorsal horn GABA levels after the onset of stimulation.5 Additionally, it was noted that the animals that failed to normalize withdrawal thresholds in response to spinal cord stimulation did not have the enhanced GABA release in the dorsal horn seen in the animals that did respond to spinal cord stimulation. This prompted further study of the nonresponding animals to determine whether the use of small doses of the GABA agonist baclofen, in conjunction with spinal cord stimulation, would alter the effectiveness of the stimulation.6 The doses of baclofen used in the study were well below those required to clinically affect withdrawal thresholds when used alone. The investigators found that the combined effects of spinal cord stimulation and low-dose baclofen resulted in diminished pain behavior in the previously nonresponding animals.

The concept of spinal cord stimulation modulating pain transmission through the effects on GABA and excitatory amino acid levels in the dorsal horn has been reinforced by a study that demonstrated that excitatory amino acid content of the dorsal horn can be diminshed during spinal cord stimulation.7 Additional research into the role of GABA in mediating the effects of spinal cord stimulation has involved the use of GABA antagonists. The administration of the GABA antagonist Bicuculline in cats subjected to spinal cord stimulation attenuated the inhibitory effects of spinal cord stimulation on transmission of ascending impulses, as measured by a recording electrode situated over the anterolateral funiculus.8 Thus, GABA-related functions play a critical role in the modulation of painful inputs, and it appears that spinal cord stimulation acts, in part, to alter dysfunctional GABA systems that may be present in chronic pain states.

Adenosine

The distribution of adenosine receptors in the dorsal horn of the spinal cord has led several investigators to speculate that adenosine, like GABA, may have important modulatory effects on the transmission of painful impulses. Experiments that parallel those used to evaluate the role of GABA in the allodynia-suppressing effect of spinal cord stimulation have been done using adenosine agonists and antagonists.9–11 These studies have shown that animals with evidence of allodynic behavior that fail to respond to spinal cord stimulation alone may become responsive to spinal cord stimulation by the addition of intrathecal adenosine agonists at clinically subeffective doses. Likewise, the intrathecal administration of small doses of an adenosine receptor antagonist to animals responsive to spinal cord stimulation resulted in apparent diminution in the effectiveness of the spinal cord stimulation in abolishing allodynic behavior.

Serotonin and Substance P

Several investigators have looked at substance P and serotonin release in the dorsal horn as a consequence of spinal cord stimulation. It is thought that these mediators have an inhibiting modulatory effect on transmission of painful stimuli in the dorsal horn. It also appears that the release of these mediators during spinal cord stimulation may, in part, occur through a supraspinal loop.12

Calcitonin Gene-Related Peptide

Calcitonin gene-related peptide (CGRP) is another putative mediator of some of the effects of spinal cord stimulation that has been investigated. CGRP release results in a vasodilatory response, which may account for some of the beneficial effects spinal cord stimulation has been shown to have on such conditions as reflex sympathetic dystrophy (complex regional pain syndrome [CRPS I]), intractable angina, and peripheral vascular disease, and disorders in which there is a significant component of ischemia. This seems to occur by antidromic stimulation of the dorsal roots, ultimately resulting in peripheral release of CGRP and subsequent vasodilation.13 These effects may occur through a nitric-oxide-dependent mechanism.

Nitric Oxide

As mentioned previously, it appears that nitric oxide may play some role in mediating the vasodilatory effects of spinal cord stimulation. Investigators tested this hypothesis by administering a nitric oxide synthase inhibitor in an animal model after the application of dorsal column stimulation. Vasodilation of the ipsilateral hindpaw seen during dorsal column stimulation was significantly diminished after the nitric oxide synthase inhibitor was given. Further study was done to evaluate whether these effects are mediated through the autonomic nervous system. The results seemed to indicate that the nitric oxide release may occur independently of the sympathetic nervous system outflow, since the administration of the ganglionic-blocking agent hexamethonium in an animal model did not diminish the increased cutaneous blood flow seen during dorsal column stimulation.14

Sympathetic Nervous System Influences

Although some of the vasodilatory effects of spinal cord stimulation appear to occur by nitric oxide and CGRP-related mechanisms, there may be some sympathetically mediated vasodilation as well. Linderoth et al examined which receptor types might be involved.15 They administered several types of receptor antagonists to further elucidate the role of cholinergic and adrenergic activity in mediating the vasodilatory effects of spinal cord stimulation. With the administration of α1-receptor antagonists, beta blockers, nicotinic-receptor antagonists, and the nonspecific ganglionic blocker hexamethonium, the vasodilatory effects of spinal cord stimulation were either diminished or completely abolished. Muscarinic receptor antagonists did not have these same effects. Thus, it appears that sympathetically mediated vasodilation in spinal cord stimulation occurs by stimulation of nicotinic- and adrenergic-receptor subtypes.

While many of the mechanisms by which spinal cord stimulation exerts its beneficial effects on painful syndromes remain under investigation, the scope of conditions that are being treated with this modality is expanding. To ensure maximal efficacy of this technology, proper patient selection must be undertaken prior to implantation of these devices. This selection process depends not only on the individual to be treated but also on the condition for which it is prescribed. If all relevant information is not reviewed prior to implantation of the spinal cord stimulator, the overall chance of success with this device may be diminished. Given the increasing demand for cost containment in the current era of medical practice, the placement of such a device in an improperly screened patient is likely to result not only in patient dissatisfaction, but also in stricter regulations by third-party payers with regard to selection of appropriate candidates for these devices. The suitability of spinal cord stimulation in any given patient depends first and foremost on the condition to be treated. The most common applications of spinal cord stimulation are in peripheral vascular disease, failed back surgery syndrome, complex regional pain syndrome, and angina pectoris.

Peripheral Vascular Disease

Peripheral vascular disease ranks among the most common indications for spinal cord stimulation worldwide. Patients with advanced peripheral vascular disease that is not amenable to revascularization procedures may be candidates for treatment with spinal cord stimulation. Also, patients with other diseases resulting in limb ischemia, such as Raynaud’s disease, Buerger’s disease, and CRPS I may derive benefit from this treatment. Intractable pain due to peripheral vascular disease generally seems to respond well to spinal cord stimulation. Additionally, it has been noted in a number of cases that healing of ischemic ulcers appeared to be enhanced in patients in whom this modality was employed. This finding prompted further study of the use of spinal cord stimulation in peripheral vascular disease with critical limb ischemia. When groups of peripheral vascular disease patients were followed over time, those who were using spinal cord stimulation as a means of treating their pain were seen to have an increased limb salvage rate as compared with their medically treated cohort.16,17 This effect is thought to be related to vasodilation mediated by the sympathetic nervous system during spinal cord stimulation, creating an enhanced microcirculation. The quality of life of these patients can be significantly improved through the use of spinal cord stimulation. This may not only result from decreased analgesic requirements but also from potentially delayed amputation of the affected limb. Thus, spinal cord stimulation may act to postpone disabling but necessary treatment in these individuals, in addition to treating their pain.

Failed Back Surgery Syndrome

Many patients who have undergone surgical procedures such as laminectomies to treat back pain and radiculopathy continue to experience pain postoperatively or redevelop pain at a later date. These patients may, in some cases, undergo one or more repeat operations to further treat their pain. Despite surgical interventions, a percentage of these patients will continue to be plagued by pain. When surgically remediable anatomic lesions such as herniated disks, spinal stenosis, or a suboptimal fusion have been corrected, patients may be considered for placement of a spinal cord stimulator. In addition to pain relief, outcome measures such as decreased use of analgesics, increased physical activity, loss of neurologic function, and return of the previously disabled to work have been examined in a number of studies.18–20 These studies showed that patients who derived pain relief from spinal cord stimulation often diminished their analgesic use or discontinued it altogether. Additionally, most were be able to perform more routine daily activities, and some who had previously been unfit to work were able to gain employment. Another study has been performed to compare outcomes when patients with failed back surgery syndrome were treated with spinal cord stimulation versus re-operation.21 In an initial series of patients, the rate of crossover from re-operation to placement of a spinal cord stimulator was much greater than the reverse. These initial results suggest that, in some patients with failed back surgery syndrome and surgically correctable lesions, spinal cord stimulation may serve as a viable alternative to re-operation in further treating pain. Traditionally, pain in a monoradicular distribution has been the most amenable to treatment by neurostimulation. With larger areas of pain, it is more difficult to adequately cover all of the pain with paresthesias, a necessary prerequisite for success in using neurostimulation. Therefore, patients with more complex pain patterns, such as bilateral radiculopathy with axial low back pain, have often been seen as less apt to respond to spinal cord stimulation. As the technology becomes more advanced, with the use of multielectrode, multichannel, and multiprogram devices, these more complex pain patterns are increasingly being viewed as responsive to spinal cord stimulation.22–26

Intractable Angina Pectoris

As with patients suffering from end-stage peripheral vascular disease, patients having severe coronary syndromes that are unamenable to revascularization may find relief through spinal cord stimulation. This population of patients has often undergone multiple prior coronary angioplasties or coronary artery bypass operations and is receiving maximal medical therapy with continued recurrent angina despite aggressive treatment. Most of these patients are fairly incapacitated by their symptoms and fit functional classifications III to IV of the New York Heart Association. The use of spinal cord stimulation in this group of patients with severe coronary artery disease serves not only to diminish their pain but appears to overall decrease the number of ischemic episodes as well. This is thought to be due, in large part, to a reduction of sympathetic nervous system activity induced by spinal cord stimulation. A number of studies have been done to evaluate the impact of spinal cord stimulation on myocardial ischemia. Generally, these studies have revealed that use of spinal cord stimulation in cases of intractable angina pectoris may lead to reduced episodes of angina and decreased requirements for short-acting nitrates.27,28 It also appears that exercise tolerance is enhanced in these patients, likely due to alterations in myocardial oxygen supply-to-demand ratio and possibly through redistribution of coronary blood flow to ischemic areas. Several studies have used techniques of atrial pacing to determine how anginal thresholds are affected by spinal cord stimulation.29–31 During these studies, patients were paced to angina before and after turning on their spinal cord stimulators. After initiating stimulation, these patients were able to tolerate pacing to higher heart rates than prior to stimulation. It was noted that these patients would ultimately still experience angina when being paced maximally. These findings are of considerable importance in this arena, as some critics have expressed concern that spinal cord stimulation may only mask the symptoms of ischemia, thereby eliminating a key warning signal for these patients. Overall, the use of spinal cord stimulation in intractable angina pectoris appears to have beneficial effects that go beyond pain relief. It may reduce the frequency of angina (i.e., ischemia) in these patients through an alteration in the dynamics of myocardial oxygen supply and demand. This may have a significant impact on the quality of life of those suffering from this disabling morbidity.32 Given the ever-increasing number of individuals in modern society who are developing coronary artery disease, partly as a consequence of lifestyle influences, it is likely that the number of patients treated with spinal cord stimulation for refractory angina will expand in the future.

Complex Regional Pain Syndrome

Complex regional pain syndrome I, previously referred to as reflex sympathetic dystrophy, is an entity that remains somewhat of an enigma in terms of its causation. As implied by its former name, this condition appears to involve aberrant activity in the sympathetic innervation of the affected body part that results in pain and other related sequelae seen in patients with this disorder. Spinal cord stimulation has a high rate of success when used in this patient population. Many of these patients have failed to respond to more conservative measures, such as sympathetic blocks and medications, and some have failed to have full response to surgical sympathectomy. Some authors have suggested that spinal cord stimulation is preferable to surgical sympathectomy in these cases.33 Retrospective reviews of a limited number of patients having spinal cord stimulators implanted for management of their complex regional pain syndrome have shown that patient satisfaction with this mode of therapy is quite high, with up to 90% of these individuals finding the stimulator helpful for their pain.34,35

Many patients may be able to discontinue opiate therapy after initiation of spinal cord stimulation. The effectiveness of this modality in treating CRPS likely relates directly to its apparent inhibition of sympathetic activity in the area of stimulation, as has been seen in other patient populations such as those with peripheral vascular disease. Control of pain in these circumstances is an important component of proper rehabilitation because patients may be significantly restricted when participating in physical or occupational therapy due to their pain. This may create a vicious cycle in which disuse of the involved extremity eventually results in irreversible atrophic changes and significant disability. Complex regional pain syndrome remains an underdiagnosed source of pain, which appears to respond favorably to spinal cord stimulation. As this condition becomes more widely recognized and understood, the application of stimulators for this phenomenon will certainly expand.

Emerging Applications

Spinal cord stimulation has been used in the management of a number of conditions aside from those previously described. These entities include phantom limb pain, post-amputation stump pain, postherpetic neuralgia, multiple sclerosis, spinal cord injury, peripheral neuropathy, and post-herniorrhaphy pain.36 The rate of permanent implantation in some of these patient populations remains low, and thus the long-term efficacy of spinal cord stimulation has yet to be established in these cases.

Patients who have an established pain diagnosis that has not responded to conservative therapies, such as medication trials and injections, may be potential candidates for spinal cord stimulation. Patients who are to be considered for possible implantation of spinal cord stimulators must undergo screening to determine appropriateness of this therapy for their conditions.37 Although some individuals may technically fit diagnostic categories suitable for spinal cord stimulation, they may ultimately be considered inappropriate candidates for a variety of physical and psychological reasons. To ensure a high likelihood of success with these devices, patients should be evaluated for significant psychopathology by a qualified psychologist or psychiatrist. During this time, patients will be evaluated for the presence of significant depression, anxiety, and personality disorders. A history of drug abuse will also be explored. If any significant psychopathology is present, or if a substance abuse problem is noted, it must be determined whether these conditions are active or controlled. Any ongoing severe psychological conditions or substance abuse must be treated prior to consideration for a spinal cord stimulation trial. Additionally, there are a number of diseases and other conditions that contraindicate implantation of these devices. Absolute and relative contraindications aside from the above include concomitant pregnancy, the presence of implanted pacemakers or automatic implantable cardioverter-defibrillator devices (AICD), malignancy as the source of pain, life expectancy of less than 1 year, inability to comprehend use of the device, severe cardiac compromise, and concomitant anticoagulation therapy. Once a patient has been appropriately screened and counseled regarding expectations, a trial of spinal cord stimulation should be arranged as a prerequisite to permanent implantation.

A neuromodulation system consists of a contact electrode or electrodes and either a subcutaneous pulse receiver or an implanted pulse generator. Electrodes are either percutaneously implantable through a Tuohy needle or through direct surgical access, for instance, in the spinal canal by a laminotomy in which the yellow ligament is removed. Electrode systems vary from a simple percutaneous four-contact electrode to complex laminectomy-style electrodes with up to 16-electrode contacts. A subcutaneous pulse receiver with an external power unit or an implantable battery unit allow for variation in stimulation parameters such as amplitude, frequency, and pulse duration. Current equipment is manufactured by Medtronic, Inc. (Minneapolis, Minn) or Advanced Neuromodulation Systems, Inc (Allen, Tex).

A fully implanted system powered with a subcutaneous pulse generator offers higher convenience to the patient. A percutaneously powered system operating through a subcutaneous pulse receiver, which is powered by a stimulator unit that is glued to the skin, offers the advantage of broader stimulation functions and obviates the need for battery replacement. Choice of system can, in part, be influenced by information obtained during trial stimulation. If energy requirements during trial stimulation are low, an implanted power source may be a reasonable choice, whereas substantial energy requirement might lead to the choice of a percutaneously powered system.

Clinical evaluation including psychological assessment helps identify patients that are suitable candidates for invasive neuromodulation therapy.38 The preoperative evaluation does not, however, provide a guarantee for long-term effect from neuromodulation in general and spinal cord stimulation for instance in particular. Since these treatments are nonablative, a trial can therefore conveniently be performed. Trial electrodes are inserted percutaneously into the epidural space of the spinal canal. This is typically done with the patient in the prone position using fluoroscopic guidance. A strict percutaneous technique can be performed where the electrode is taped or sutured to the skin. The advantage of this approach is that the trial electrode can be removed during an office visit. If the trial leads to permanent implantation, a new electrode system will have to be implanted. Another approach is anchoring the test electrode to the superficial fascia and connecting this electrode to a percutaneous extension. If the trial is positive, this allows for keeping the trial electrode as the permanent stimulating electrode. Either one or more electrodes can be positioned during the trial procedure. The trial implantation is performed as a same-day-surgery procedure. Care should be taken to minimize tissue manipulation so that post-procedural pain does not interfere with interpretation of the test results. An unintended dural perforation can for instance also obscure the test results if significant CSF leakage occurs over the following days creating a spinal headache. The trial electrode implantation is done under local anesthesia with light intravenous sedation. When the electrode is in place, intraoperative test stimulation is performed. It is important that the patient is able to clearly understand questions and communicate with the surgeon during this part of the procedure. The subjective sensation of stimulation parasthesias, induced while stimulating, has to overlap the area where pain is felt in order to obtain pain relief.39 After insertion of trial electrodes the patient is carefully instructed on how to perform the trial. The patient should keep a log during the trial in which pain intensity is documented when stimulating, as well as when not stimulating. In addition, consumption of pain medication and activity levels should be monitored. The trial period should be of a duration that clearly allows the patient to get a sense of the effect of stimulation. Three to eight days is a reasonable duration. Reporting at least 50% reduction in pain during the trial period without an increase in pain medication, or even better with a reduction in pain medication, is considered a positive trial and will lead to implantation of a permanent device.

A positive test trial, however, is no guarantee that spinal cord stimulation will provide long-term useful pain relief. A systematic review of the 1995 literature19 indicated that an average of 59% of patients had greater than 50% pain relief at a mean follow-up of 16 months. Even at centers with significant experience, up to half of patients initially implanted with permanent spinal cord stimulator systems will experience long-term failure.

A number of factors influence the choice of permanent spinal cord stimulation hardware. The choice often narrows down to whether to use a percutaneous electrode system or a laminectomy-style system for permanent stimulation. The laminectomy system requires removal of the test stimulation electrode and a direct surgical exposure of the spine. Subsequently, a laminotomy is performed, and through this spinal window, the laminectomy-style electrode is inserted. This bigger initial effort, however, is rewarded by a lower energy need for same effect, which translates into longer battery life. There is less tendency of the laminectomy-style electrode to migrate after implantation. The stimulation coverage is larger. There is less of a tendency of the stimulation to vary depending on body position. Lastly, a not infrequently seen side effect of stimulation-associated back pain is eliminated. The stimulation-associated back pain is likely to be the result of direct electrical stimulation of nociceptive fibers in the yellow ligament when using the percutaneous electrodes. The laminectomy-style electrode has a dorsal insulation, preventing stimulation of dorsally located structures.40,41 In the cervical spine it is, however, not uncommon to have good stimulation coverage at low-energy requirement when using a percutaneous electrode. A percutaneous electrode may therefore in this setting be an appropriate electrode choice.

Permanent electrode placement using either percutaneous technique or laminectomy-style technique is again performed with the patient in a prone position resting on a fluoroscopic table. A lateral or semilateral position can be used, but anatomic interpretation of landmarks as well as fluoroscopic imagery is greatly facilitated when having the patient in a strict prone position. Using local anesthesia supplemented with intravenous sedation, any point of the spinal canal from occiput to lumbar region can be accessed. The lead placement includes fluoroscopic guidance as well as intraoperative test stimulation with patient feedback. When the proper stimulation pattern has been achieved, the lead should be secured to either yellow ligament or fascia to prevent postoperative migration.

A number of locations are possible for the implantable pulse generator or subcutaneous pulse receiver. These include lateral chest wall, abdomen, infraclavicular area, and upper gluteal area. An abdominal site often necessitates a two-step procedure; first the electrode is inserted with the patient in a prone position and, subsequently, tunneling and placement of pulse receiver or pulse generator in a lateral position. The softness of abdominal wall tissue can also pose difficulties for adequate anchoring of the subcutaneous device. The lateral chest wall can often be reached with the patient in a prone position. Placing the device over the lateral chest invariably eliminates the possibility for sleeping on the side of the implant. The pectoral or infraclavicular region can be quite suitable but may pose cosmetic issues. The gluteal area is in most instances the most suitable location. The exact positioning is important and should in general be centered approximately 7 to 8 cm below the iliac crest.

The weakest mechanical link in a spinal cord stimulation system is the electrode. Electrode migration can occur, shifting stimulation coverage. As previously indicated, a laminectomy electrode reduces this risk. The increased mobility of the cervical spine increases the risk of electrode fracture secondary to metal fatigue given the amplitude of cervical spine movements. Intraspinal CSF leak following accidental dural puncture leading to spinal headache can in rare instances necessitate an epidural blood patch. Serious neurologic complications such as weakness or paralysis occur rarely. Epidural fibrosis formed around previously placed epidural stimulator electrodes can pose an additional challenge when performing revisions of such systems. Multilevel laminotomy, or even partial laminectomy, will reduce the risk of injury to the spinal cord and epidural bleeding, which otherwise can be the risk from blind sublaminar dissection of a fibrosed epidural space.

Intracranial Stimulation

Intracranial stimulation has focused on deep structures in the brain, and more recently, on the stimulation of cortical structures. Thalamic stimulation for pain was first reported in 1960.42 Stimulation of the ventropostero lateral (VPL) nucleus of the thalamus provided relief from chronic intractable neuropathic pain. A similar effect was subsequently reported from chronic ventroposterior-medial (VPM) stimulation.43 Stimulation of VPL and VPM is like spinal cord stimulation associated with a sensation of parasthesias of the target area. Stimulation of periaqueiductal gray matter (PAG) and periventricular gray matter (PVG)44–46 is more efficient in treating nociceptive pain. PAG and PVG stimulation is reversible by naloxone, and therefore in part mediated by endogenous opioids. Studies in chronic pain patients have shown increased spontaneous abnormal firing pattern in VPL/VPM.47,48 Electrical stimulation in these areas may conceivably block the abnormal bursting activity, leading to the observed pain relief. Long-term success of deep brain stimulation for nociceptive pain is in the middle 60%, whereas the effect for neuropathic pain is in the low 40%.

Implantation of deep brain stimulating electrodes require precise placement of the stimulating electrode in the intended target. The procedure is done under local anesthesia, during which a stereotactic frame is attached to the skull of the patient. Subsequently, a computed tomography (CT) scan or magnetic resonance imaging (MRI) scan is obtained with a localizing frame attached to the stereotactic base frame. The location of the anterior and posterior commeasure is determined with reference to the stereotactic frame. Assisted by a stereotactic atlas of the brain, the surgeon can with this information, locate the intended anatomic target. Intraoperatively, the anatomic target is confirmed through physiologic localization. This can include microelectrode recordings as well as micro- and macrostimulation, particularly in the thalamic targets. Stimulation in PVG or PAG is commonly associated with changes in blood pressure and heart rate. For localization of the later target, the surgeon must, however, rely primarily on anatomic landmarks. The permanent deep-brain-stimulating electrode is a quadipolar electrode (Medtronic, Minneapolis, Minn). The electrode is hooked up to a subcutaneously placed pulse generator that is placed in the subclavicular region. Since PVG/PAG stimulation is in part mediated via endogenous opioides, it is often possible to get bilateral pain relief from a single electrode inserted in the nondominant hemisphere. For unilateral pain, a contralateral PAG/PVG electrode is inserted. For VPM/VPL stimulation, effective relief of pain is only achieved from an electrode implanted in the contralateral hemisphere.

The observation of hyperactive bursting neuronal activity central to lesions in the somatosensory system prompted the concept of cortical stimulation. Stimulation of the precentral motor cortex turned out to provide pain relief as opposed to stimulation of the post-central cortex in patients experiencing post-stroke thalamic pain.49,50 Motor cortex stimulation has also been used in the treatment of trigeminal neuropathic pain.51 A laminectomy-style electrode is implanted into the epidural space during surgery. The location of the motor cortex is determined on the basis of anatomic landmarks as well as electrophysiologic testing. The electrode is thereafter exteriorized through a separate stab incision and test stimulation is performed for up to a week. If effective, an implantable battery unit is subcutaneously implanted in the subclavicular region and connected to the epidural lead through a subcutaneous extension. A prerequisite for effect is presence of a relatively intact motor system on clinical exam. In a study by Yamamoto et al, more than 70% of patients reported excellent pain relief by chronic motor cortex stimulation.52 Motor cortex stimulation shows promise in treatment of these otherwise treatment-resistant conditions.